557 
hyoid arch and it remains undifferentiated from the columella proper 
until development is well advanced. Furthermore, if its processus 
ventralis represents the interhyale, or at least the rudiment of it, 
as I think it does, this alone seems an adequate reason for considering 
the extra-columella a primary element. Because without it, the 
primitive condition could not be established inasmuch as the inter- 
hyale is connected to it, and not to the columella proper. 
Another point of difference between Testudo graeca and Chrys- 
emys marginata is in the formation of the tympanic cavity. In this 
stage the hollow of the quadrate has extended anteriorly and poste- 
riorly and reminds one of the adult condition. The posterior portion 
of the quadrate is still in a partial precartilage stage, and very close 
to its dorso-lateral edge, the squamosal is appearing as a thin shell 
of bone. The distal ends of the visceral pouch and the extra-colu- 
mella have shifted somewhat anteriorly, so that the ventral portion 
of the cavity, which has widened out into a small pouch, nearly fills 
the anterior region of the hollow, and the dorsal tip now lies well 
within the hollow also, and shows a tendency to expand into the 
posterior part. The extra-columella, instead of occupying the poste- 
rior margin of the hollow as formerly, has swung into a latero-median 
position. The shaft of the columella is nearly surrounded by the 
tympanic cavity, dorsally by the smaller posteriorly directed portion 
and ventrally by the large inferior part at the point where the median 
division (the tympanic tube) opens into it. This median portion passes 
anteriorly between the quadrate and otic capsule and opens broadly 
into the pharynx. Thus the formation of the tympanic cavity and 
tube in Chrysemys marginata is much simpler than that of Testudo 
graeca where BENDER found both the first and second visceral pouches 
contributing to its formation. The condition in Chrysemys is more 
like that of Emys europaea which Noack describes as a derivative 
of the first pouch. BENDER is inclined to think that on account of 
young embryos Noack may have made a misinterpretation and that 
the second cleft may have entered into the formation of the tympanum. 
My own results drawn from a wide range of closely intergrading series 
(which show no suggestion of the second visceral cleft playing any 
part in the formation of this structure) — lead me to think that 
Noacx’s own interpretation is correct. Furthermore, Corps, who 
has done the most extensive work on the reptilian tympanic cavity, 
describes in Lacerta agilis a development and final formation which 
